Benefits versus risks of pharmacological prophylaxis to prevent symptomatic venous thromboembolism in unselected medical patients revisited. Meta-analysis of the medical literature.

A significant proportion of the outcomes reported in trials assessing venous thromboembolism (VTE) prophylaxis in medical patients are related to asymptomatic events found on routine imaging studies. The implications of these events are controversial. Moreover, such trials did not always reflect the patient mix in today's internal medicine departments. We summarized the evidence assessing the rate of symptomatic VTE events and the benefit of pharmacological prophylaxis in unselected medical patients, and formally evaluated the benefit versus risk of this intervention. We searched MEDLINE, EMBASE and CENTRAL until June 2011 for studies that prospectively followed cohorts of medical patients and assessed the rates of VTE, and randomized controlled trials reporting the effect of prophylaxis on these events, at 3 weeks and 3 months. Eight trials were included. The rates of symptomatic VTE were 0.69 and 3.7 % for short term and long term follow-up periods, respectively. In the interventional meta-analysis, the odds ratio (OR) for overall mortality and for symptomatic VTE at 3 weeks were 0.93 and 0.59, favouring intervention. The OR for major bleeding at 3 weeks was 2.0, favouring no intervention. None of these results were statistically significant. The number needed to treat to prevent one overt VTE event was 292, while the number needed to treat for an additional major bleeding was 336. In unselected medical patients, the rate of symptomatic VTE is lower than the reported overall VTE rate, and the benefit to risk ratio of pharmacological intervention for alleviating this condition in at-risk medical inpatient is questionable. Further specifying the population at risk for an overt VTE, and the clinical significance of asymptomatic events, is warranted.